VideoMAE_BdSLW60_100_0.15splt_gradAcu

This model is a fine-tuned version of MCG-NJU/videomae-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.9249
• Accuracy: 0.8323

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 22400

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
3.771	0.0400	897	3.3314	0.1894
1.3109	1.0401	1795	1.0711	0.7706
0.5765	2.0401	2693	0.7444	0.8012
0.4197	3.0401	3591	0.6252	0.82
0.273	4.0400	4488	0.3732	0.8965
0.2655	5.0401	5386	0.4039	0.9
0.2046	6.0401	6284	0.3899	0.9024
0.1659	7.0401	7182	0.3334	0.9212
0.1378	8.0400	8079	0.3984	0.9059
0.1599	9.0401	8977	0.1934	0.9518
0.1178	10.0401	9875	0.3126	0.9259
0.1445	11.0401	10773	0.1771	0.9565
0.0843	12.0400	11670	0.3163	0.9259
0.0792	13.0401	12568	0.3593	0.9235
0.0376	14.0401	13466	0.2353	0.9447
0.0817	15.0401	14364	0.4593	0.9047
0.1005	16.0400	15261	0.1929	0.9576
0.0585	17.0401	16159	0.1758	0.9635
0.0407	18.0401	17057	0.1733	0.9635
0.088	19.0401	17955	0.1781	0.9647
0.0489	20.0400	18852	0.1547	0.9659
0.0514	21.0401	19750	0.1564	0.9682
0.0246	22.0401	20648	0.1298	0.9694
0.0445	23.0401	21546	0.1642	0.9694
0.017	24.0381	22400	0.1573	0.9671

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1